PROJECT SUMMARY Immune checkpoint inhibitors (ICI), e.g., anti-CTLA-4 (ipilimumab) or anti-PD-1 (nivolumab, pembrolizumab), have transformed the therapeutic landscape for advanced melanoma, enhancing anti-tumor activity and producing durable clinical benefit in a subset of patients. ICIs have been shown to be effective in the adjuvant setting for melanoma, and for treatment of many other cancers (e.g., bladder, lung, renal cell). However, immune-related adverse events (irAEs) are a critical obstacle to realizing the full potential of ICI: a substantial proportion of ICI-treated patients develop severe immune toxicities involving multiple organs, causing significant morbidity and requiring systemic treatment or therapy discontinuation. In pilot testing, we used a human proteome array with ~20,000 full-length human proteins to analyze antibody levels in sera from 67 melanoma patients who received anti-CTLA-4 or anti-PD-1. Our results identified distinct pre-treatment serum autoantibodies (autoAbs) associated with development of severe irAEs. Using a humanized FcgR mouse model, we generated preliminary data to suggest that pre-treatment patient sera IgG may exacerbate development of irAEs. Finally, using a novel mass spectrometry approach we demonstrated an association between specific serum proteins from ICI-treated melanoma patients and treatment efficacy. Our central hypothesis is that a subset of melanoma patients has a baseline autoimmune susceptibility, characterized by a repertoire of specific preexisting autoAbs and serum proteins that predicts and exacerbates development of ICI toxicity. Our proposal is in direct response to RFA-CA-17-017 - PQ8: ?What are the predictive biomarkers for the onset of immune-related adverse events associated with checkpoint inhibition, and are they related to markers for efficacy?? We will utilize a large cohort of pre-treatment sera from the CheckMate-238 phase 3 trial of adjuvant anti-CTLA-4 vs. anti-PD-1 in resected Stage III/IV melanoma to assess the relationship between autoAbs and serum proteins and development of irAEs and recurrence-free survival (efficacy). We will use our mouse model to test the cause-effect relationship between pre-existing immune responses and toxicities from ICI to select the autoAbs most likely to be biomarkers for irAEs. The translational impact of our work is that detection of baseline toxicity-associated autoAbs could identify melanoma patients likely to develop severe irAEs from treatment, guiding therapy selection and sequencing or toxicity management. Our work with autoAbs and serum proteomics may define new targets which impact the onset of irAEs and provide insight into new strategies to mitigate toxicity without compromising the anti-tumor immune response.